J . Med. Chem. 1994,37, 3171-3180
3171
Decapeptide Agonists of Human C5a: The Relationship between Conformation and Spasmogenic and Platelet Aggregatory Activities? Sam D. Sanderson,*’$Leonid Kirnarsky,$ Simon A. Sherman,* Julia A. Ember,$ Angela M. Finch,l and Stephen M. Tay1or-L Eppley Institute for Research in Cancer and Allied Diseases, University of Nebraska Medical Center, 600 South 42nd Street, Omaha, Nebraska 68198-6805, Department of Immunology, Scripps Research Institute, 10666 North Torrey Pines Road, La Jolla, California 92037, and Department of Physiology and Pharmacology, University of Queensland, St. Lucia, QLD 4072, Australia Received April 22, 1994@
A series of decapeptide analogues corresponding to the C-terminal region of human C5a anaphylatoxin (C5a65-74) was synthesized with residue substitutions to restrict conformational flexibility in the C-terminus. These conformationally constrained peptides behaved as agonists of C5a in spasmogenic assays (smooth muscle contraction in human fetal artery, guinea pig ileum, and guinea pig lung parenchyma) as well as guinea pig platelet aggregation. There were significant correlations in the potencies of these peptides between the various assays. A structure-function analysis led to the identification of a preferred backbone conformation that correlated with the expression of these biological responses. These backbone structural motifs were consistent with a helix-like conformation for residues 65-69, an elongated structure for residues 70-71, and a p-turn of either type I1 or type V for residues (71172-74. The most potent of these agonists expressed almost 5% of the potency of natural C5a.
Introduction The human anaphylatoxin C5a is a 74-residue glycopolypeptide that is generated by the enzymatic cleavage of the fifth component ((25) during plasma complement activation. C5a is recognized as a principal mediator of local and systemic inflammatory responses because of its ability to activate and recruit neutrophils,l induce spasmogenesis? increase vascular ~ermeability,3-~ and stimulate the release of secondary inflammatory mediators from a variety of cell types.6-8 C5a also appears to play a role in the modulation of immune response because of its ability to induce directly or indirectly the synthesis and release of the cytokines interleukin-1 (IL-11, interleukin-6 (IL-61,interleukin-8 (IL-8),and tumor necrosis factor-a (TNF-a) from human monocyte^.^-'^ These inflammatory and immunomodulatory activities are expressed via a transmembrane, G-protein-mediatedsignal transduction mechanism when the C5a ligand interacts with its receptor(s) expressed on the surface of certain circulating and tissue cell types. 13-16 Because of its proinflammatory activity, C5a has been implicated as a pathogenic factor in the expression of certain inflammatory disorders such as rheumatoid arthritis,17 adult respiratory distress syndrome,l8Jg gingivitis,20 and the tissue damage associated with atherosclerosis and myocardial infarction.21r22Consequently, there is considerable interest in developing a specific C5a antagonist for use as an antiinflammatory agent in the treatment of these diseases. + This work is supported by grants from the Nebraska Cancer and Smoking Disease Research Program no. 94-40to S.A.S. and no. 93-42 to S.D.S.and NCI CA3G727, ACS SIG16, an American Heart Association Fellowship to J.A.E., and an NSRG grant to S.M.T. from the University of Queensland. * Author to whom correspondence should be addressed. University of Nebraska Medical Center. + Scripps Research Institute. University of Queensland. Abstract published in Advance ACS Abstracts, August 15, 1994.
*
@
One approach to the development of a C5a antagonist has focused on the synthetic manipulation of peptides possessing sequence homology to the C-terminal “effector” region of natural C5a. These peptides have been shown to be effective agonists compared to the parent polypeptide, but at markedly reduced p o t e n ~ i e s . ’ ~ , ~ ~ , ~ ~ Therefore, a first step toward the development of an antagonist has been to increase the potency of these agonist peptides to a level approaching that of natural C5a, the rationale being that the increase in potency would reflect a heightened affinity for the C5a receptor. A potent, high affinity peptide could then serve as a template from which an analogue or mimetic might be developed that would retain the high affinity binding characteristics, so as to compete with natural C5a for the receptor, but not transduce a biological signal when bound to the receptor. This “ligand” approach has generated numerous agonist analogues that express a variety of biological activities and potencies, some of which have been used as probes of C5a pharmacol~gy.~~ However, an effective peptide-based C5a antagonist has yet to be clearly described. In order to propose a rational design strategy ’for the generation of a high affinityhigh potency template analogue, it is necessary t o elucidate the relationship between the structural features expressed by a C-terminal agonist peptide and their biological consequences. The general spatial arrangements of the N-terminal region of human C5a (residues 1-63) have been determined on the basis of lH-NMR data, but no definable spatial structure could be assigned to the C-terminal (‘effector”region (residues 64-74),26 which appeared more flexible than the rest of the C5a polypeptide. However, based on a detailed analysis of the sequential d-connectivities from the original NMR data on human C5a,26we were able to distinguish three general structural features of the C-terminal decapeptide region, C5a65-74 (ISHKDMQLGR). The region comprised of residues Ile-65 to Asp-69 was consistent with a twisted,
0022-262319411837-3171$04.50/0 0 1994 American Chemical Society
3172 Journal of Medicinal Chemistry, 1994, Vol. 37, No. 19
Sanderson et al.
helix-like structure, and the region made up of Met-70 resented by the C-terminal ten residues (C'5a65-741, we to Gln-71 possessed an elongated backbone conformaapplied a probabalistic approach29 that utilized the tion. The C-terminal region of C5a65-74, comprised of FISINOE pr~gram.~'J Input data to the FISINOE residues Leu-72 to Arg-74, was considerably more program came from the presence and/or absence of flexible than the rest of the peptide and appeared to be sequential d-connectivities from the original IH-NMR made up of overlapping structural contributions of both data on human C5a.26 twisted and elongated conformations. Flexibility in the The simultaneous presence of sequential d a - and C-terminal region (residues 71-74) appears to be imd"-connectivities in the region (64165-69 as well as portant in the expression of biological activity because d M (i, i 3) was consistent with this region of C5a dramatic changes in activity and potency have been adopting a helix-like backbone c o n f ~ r m a t i o n . ~The ~-~~ observed when the flexibility in this region was reregion 70-71 was characterized by the presence of d a stri~ted.~~,~~,~~ connectivities and the absence of d"-connectivities, Thus, in order to develop rational design strategies suggestive of an elongated backbone c o n f ~ r m a t i o n . ~ ~ - ~ ~ that utilize these C-terminal agonist peptides, it is The C-terminal end (residues 72-74) was characterized necessary t o determine the conformational features by the presence of d o - and dm-connectivities and by within the C-terminal, flexible region of these decapepthe absence of medium- and long-range NOEs, a pattern tides and to understand the role they play in the consistent with the dynamic averaging of an ensemble expression of biological activity. In this light, it is now of structures with overlapping contributions made by possible to use the backbone conformational motifs elongated and twisted backbone conformation^.^^-^^ described above for the decapeptide C5a65-74 as a frame Thus, the NMR-matched,backbone conformational feaof reference to search for the biologically active confortures of the C-terminal 10-residue region of C5a mation(s1 and to rationalize the design for more potent (C5a65-74) can be described as consisting of twisted, and selective decapeptide agonist analogues of C5a. helix-like conformation for residues 65-69, elongated conformation for residues 70-71, and flexible structure In this paper, we report the results of a structureof overlapping contributions made by twisted and function analysis of a panel of decapeptide analogues of human C5a65-75 synthesized with residue substituelongated conformations in the C-terminal region 72tions that restrict flexibility in the C-terminal region 74. of the peptide (residues 70-74). These conformationally It is assumed that the one (or more) biologically active constrained analogues were generated in order to conformation(s) of the C-terminal decapeptide C5a65-74 determine the most likely conformation(s) that are lies within a low energy ensemble of conformers generresponsible for the expression of spasmogenic activity ated by the sterically allowable range of flexibility in human fetal artery, guinea pig ileum, and guinea pig within the C-terminal end (residues 72-74). In fact, lung parenchyma and for aggregatory activity of guinea flexibility in this end of the peptide appears to be pig platelets. For the expression of these biological important for the expression of biological activity beresponses, we identified a characteristic C-terminal, cause activity is dramatically affected when this flexturn-like motif that was consistent with a @-turnof type ibility is r e s t r i ~ t e d . ~ Thus, ~ , ~ ~ it ! ~is~ important to I1 or V for the region comprising residues 71-74. identify the conformational features in the C-terminal Moreover, analogues that expressed a structural proend of C5a65-74 that relate t o the expression of various pensity to this C-terminal motif appeared particularly biological activities. potent relative to C5a in this class of biological assays. We suggest the likelihood of a reverse or @-turnin This study is a step toward the identification of distinct the C-terminal region of (Gln-71 to Arg-74) of C5as5-74 conformational features that appear t o play a unique on the basis of three lines of evidence. First was the role in the expression of particular patterns of biological presence of overlapping twisted and elongated structure responses. The characterization of these conformational in this region as determined from our analysis of the features will contribute t o the generation of pharmaNMR data of human C5a. Second was an increase in cophore models that may assist in antagonist design neutrophil membrane binding and chemotactic activity strategies and help explain the observed variability in for certain C-terminal octapeptide analogues of C5a in the biological activities expressed by C-terminal agonist which Gly-73 was substituted with D-Ala.32 This, along peptides in different cellular and tissue ~ y ~ t e m s . ~with ~ the ~ ~position ~ , ~of~the~ Gly ~ in ~ the four-residue reverseturn stretch (i 2, or -Gln-Leu-Gly-Arg),was suggesResults and Discussion tive of a @-turn that was stabilized by the D-Ma substitution. Indeed, the pattern of changes observed Conformational Characterization. On the basis in the biological activity expressed by the conformationof the solution NMR data on human C5a, the tertiary ally constrained analogues described in this paper (see structure for the region 1-63 was derived from the below) not only confirm the presence of a C-terminal, analysis of long- and medium-ranged nuclear Overreverse-turn but correlate with a p-turn of type I1 or V hauser effects (NOE).26For the C-terminal region 64for the biological responses tested (i.e., spasmogenic and 74, however, long-range NOES were not observed. Neither were there any observable NOEs between this platelet aggregatory activities). C-terminal region and other parts of the polypeptide. Peptide Characterization. Table 1 summarizes Moreover, quantitative NOE interpretation was comthe amino acid compositional and mass spectral analyplicated by spectral overlaps. This region of C5a, ses of the 23 peptide analogues used in this study. This therefore, was interpreted to be made up of flexible, panel is also listed in Table 2, which summarizes the random structure. biological results in human fetal artery and guinea pig In order t o determine the probable backbone conforplatelets. Selected peptides are shown in Table 3, which mations within the flexible region of human C5a repsummarizes the results in guinea pig ileum and lung
+
+
Journal of Medicinal Chemistry, 1994, Vol. 37, No. 19 3173
Decapeptide Agonists of Human C5a
Table 1. Amino Acid Compositional and Mass Spectral Analysis of C5a Analoeue Peptides" Peptide
Mass Specaal Results
Amino Acids
NO.
ThLQbL 1195 1195 1245 1245 1285 1285 1228 1228 1214 1214 1198 1198 1259 1259 1259 1259 1245 1245 1228 1228 1172 1172 1186 1186 1304 ND 1254 ND 1154 1154 1186 1186 1232 1232 1230 1230 1168 ND (1) 1.88 1.24 1.10
1172 ND 1154 ND
0.99 (1) 0.98
1172 ND 1156 ND
Values in parentheses are the expected number of residues, numbers directly above them are the observed. ND, not done.
parenchyma. All peptides were based on the C-terminal 10-residues of human C5a (C5a65-74 or ISHKDMQLGR) and on a more potent analogue (C5ass-~Y65,F67or YSFKDMQLGR). The replacement of His at position 67 with the aromatic residue Phe (peptide 1,Table 2) was shown t o afford about a 2 order of magnitude ~ ' ~was ~ increase in potency relative to C 5 a ~ 5 - 7 4 . ~Tyr used in place of Ile-65 in order to provide a site for radioactive iodination for tracer studies.41 In smooth muscle contraction of human fetal artery, C5a65-74Y65,F67 was shown to be about 20 times more potent than the C-terminal 19-mer of natural sequence (C5a56-74) (Figure 1). Substituted residues that differ from those in C5a65-74Y65,F67 are shown in bold face (Table 2 ) . All peptides were homogeneous by both analytical RP-HPLC and mass spectral analysis and gave the expected residue molar ratios by compositional analysis and parent ion molecular mass by FAB-MS (see Table 1). Residue substitutions in C5a65-74Y65,F67 were chosen t o restrict the flexibility in the C-terminal region of the decapeptide in order to bias certain features of backbone conformation that would be useful in the search for the biologically relevant conformation(s) in
the flexible, C-terminal region. Three principal types of modifications were employed. These include (1)Pro substitutions for restricting local q5 angle flexibility and for influencing the allowed conformations of the preproline residue, (2) Ala substitutions for evaluating the contributions made by the side chains in the peptide, and (3) D-residue substitutions for altering local stereoisometric arrangements. The changes in biological activity induced by these restrictions in C-terminal flexibility and conformational space alterations were assessed in spasmogenic assays (smooth muscle contraction in human fetal artery, guinea pig ileum, guinea pig lung parenchyma) and in guinea pig platelet aggregation. Pharmacological Activity. Table 2 summarizes the pharmacological activities of human C5a, C5a65-74Y65,F67, and its analogues in the smooth muscle contraction of human fetal artery and the aggregation of guinea pig platelets. Table 3 summarizes the smooth muscle contractile responses of C5a and selected analogues in guinea pig ileum and guinea pig lung parenchyma. All analogues screened in these spasmogenic assays induced responses in a dose-dependent manner (see Figure 1, for example) and were shown t o be full
3174 Journal of Medicinal Chemistry, 1994, Vol. 37,No. 19 Table 2. Pharmacological Activities of C5a Analogues in Human Fetal Artery and Guinea Pig Platelet Aggregation Assays
C5a 1 ISFKDMQLGR
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
YSFKDMQLGR YSFKDMQLPR YSFKDMQPGR YSFKDMPLGR YSFKDMPPGR YSFKDMQIAR YSFKDMQLaR YSFKDMQIGR Y SFKDMPLaR YSFKDMPAGR YSFKDMPM YSFKDMPLfR YSFKDMPLpR YSFKDAPLGR Y SFKDCPLGR YSFKDCPLCR
19 20 21 22 23
YSFKDAPLaR Y SFKGMLLGR YSFKGULGR YSFKGMLLGr YSFKGMPLGR
Y SFKDCPLCR
7.92 f 0.09 (23) 5.05 f 0.06 (26)
